Icelandic TV show Landinn visited Kristinn R. Thórisson at IIIM headquarters in Nauthólsvík to gain insight into the Humanobs project. Dr. The international research group, led by Dr. Thórisson, developed a computer that can learn continuously in real-time, replicating human learning as closely as possible.
Dr. Julius Togelius gave a talk on the challenges of general video game playing at IIIM’s and CADIA’s AI Festival. AI agents have been developed for a large number of games, but being able to play a single game well does not mean that an agent possesses any general intelligence. An agent that plays Pac-Man well could probably not play Super Mario Bros. A human player would, after learning to play 2D arcade games in general, quickly learn the specifics of each game. Continue reading The Challenges of General Video Game Playing (Video from AI Festival 2014)→
Dr. Stephan Schiffel gave a talk on General Game Playing at IIIM’s and CADIA’s AI Festival. In the last decades, Artificial Intelligence was very successful in developing programs that play on par with human experts or even beat them in such complex games as Chess, Checkers, Backgammon or Poker. Why are there still other games we have not solved – for example the board-game Go? The area of General Game Playing tries to bridge the gap between these different games and develop more general solutions. Dr. Schiffel provides a look behind the success of CADIA in this field and discuss how this research is relevant beyond playing games. Continue reading Playing to Win: Success of CADIA´s General Game Playing Machine (Video from AI Festival 2014)→
Dr. Claes Strannegård gave a talk at the last AI Festival, focusing on AI beyond one trick programs — describing and demonstrating a system that mimics certain aspects of children’s cognitive development, maintaining versatility.